Critical Role of Water and Oxygen Defects in CâO Scission during

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

Article

Critical Role of Water and Oxygen Defects in C– O Scission during CO2 Reduction on Zn2GeO4 (010) Jing Yang, Yanlu Li, Xian Zhao, and Weiliu Fan Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b03360 • Publication Date (Web): 01 Mar 2018 Downloaded from http://pubs.acs.org on March 3, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Critical Role of Water and Oxygen Defects in C–O Scission during CO2 Reduction on Zn2GeO4 (010) Jing Yang†, Yanlu Li‡, Xian Zhao‡, Weiliu Fan†*

†

Key Laboratory for Colloid and Interface Chemistry of State Educating Ministry, School of

Chemistry and Chemical Engineering, Shandong University, Jinan 250100 China ‡State Key Laboratory of Crystal Materials, Shandong University, Jinan 250100 China

KEYWORDS: C-O bond scission, H2O effect, oxygen vacancy, density functional theory

ACS Paragon Plus Environment

1

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 51

ABSTRACT: Exploration of catalyst structure and environmental sensitivity for C-O bond scission is essential for improving the conversion efficiency due to the inertness of CO2. We performed density functional theory calculations to understand the influence of the properties of adsorbed water and the reciprocal action with oxygen vacancy on the CO2 dissociation mechanism on a Zn2GeO4 (010). When a perfect surface was hydrated, the introduction of H2O was predicted to promote the scission step by two modes based on its appearance, with the greatest enhancement from dissociative adsorbed H2O. The dissociative H2O lowers the barrier and reaction energy of CO2 dissociation through hydrogen bonding to preactivate the C-O bond and assisted scission via a COOH intermediate. The perfect surface with bidentate-binding H2O was energetically more favorable for CO2 dissociation than the surface with monodentatebinding H2O. Direct dissociation was energetically favored by the former, while monodentate H2O facilitated the H-assisted pathway. The defective surface exhibited a higher reactivity for CO2 decomposition than the perfect surface because the generation of oxygen vacancies could disperse the product location. When the defective surface was hydrated, the reciprocal action for vacancy and surface H2O on CO2 dissociation was related to the vacancy type. The presence of H2O substantially decreased the reaction energy for direct dissociation of CO2 on O2c1- and O3c2defect surfaces, which converting the endoergic reaction to an exoergic reaction. However, the increased decomposition barrier made the step kinetically unfavorable and reduced the reaction rate. When H2O was present on the O2c2-defect surface, both the barrier and reaction energy for direct dissociation were invariable. This result indicated that the introduction of H2O had little effect on the kinetics and thermodynamics. Moreover, the H-assisted pathway was suppressed on all hydrated defect surfaces. These results provide a theoretical perspective for the design of highly efficient catalysts.


2

Page 3 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

INTRODUCTION Recycling of CO2 into high value-added fuels will be environmentally beneficial and promote economic development. Several conversion methods can be used to recycle CO2, including thermocatalytic,1, 2 photocatalytic,3 and electrocatalytic processes.4, 5 During these processes, the activation of the C–O bond is considered a crucial step with a high energy barrier.6, 7 However, the mechanism of C–O bond cleavage remains controversial. The experimental results of Behm et al.8 and theoretical works of Zapol et al.9 suggested that CO2 was initially converted to CO by direct cleavage of the C–O bond. Alternatively, an appealing COOH precursor of C–O scission is favored.10, 11 Rodriguez et al. detected an unstable COOH intermediate by performing infrared absorption spectroscopy that indicated a CO↔OH interaction.12 These results revealed that CO2 activation is a complex process and it is necessary to clarify the dependence of C–O bond breakage on the nature of the catalyst. Zn2GeO4, as an emerging mixed-metal-oxide, is highly efficient for CO2 reduction.13,

14

However, the activation mechanism is unknown. Therefore,

characterization of the C–O bond dissociation mechanism on the Zn2GeO4 catalyst is integral to understanding the CO2 reduction process. Currently, the reaction rates and selectivity of a heterogeneous catalytic process are maximized by controlling the surface state with different structures and environments.15-17 Water is ubiquitous in surface environments and should induce the surface rearrangement and electronic reassignment of atoms to regulate the nature of active sites.18,

19

Surface catalysis

studies recently determined that the presence of water on TiO2 substantially alter the reactant or product adsorption behavior and influenced the catalytic performance.20, 21 It has been proposed that water provides H to protonate reactants.22, 23 Ge et al. reported that a COOH species was formed by CO2 and surface water on β-Ga2O3, which suggested that water provides a feasible


3

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 51

route to C–O bond breaking via a COOH precursor.24 Recent experimental studies by Zou et al. noted that the presence of water vapor substantially improved the reduction of CO2 into CH4 on a Zn2GeO4 catalyst.25 Theoretical studies have not yet clarified the origin of this surface-mediated promotion. A theoretical interpretation of the relationship between the surface environment and reaction activity is crucial to the development of highly efficient catalysts with infusive activity, especially with the features of adsorbed H2O. Surface moisture is pivotal to heterogeneous catalysis, which originates from the various properties of water at different annealing temperatures.26-28 Kwak et al. demonstrated by FTIR that the carbonate and linearly adsorbed CO2 replaced some bicarbonate when the calcination temperature was increased to 573 K. This result indicated that the appearance of CO2 on the catalyst surface was related to the hydroxyl and adsorbed H2O content.27 Chandler recently suggested that one to two monolayers of water could maximize the activity for CO preferential oxide on the catalyst surface.28 A detailed analysis that relates the surface characteristics and water reaction activity is needed to understand CO2 reduction. Surface defects that are frequently introduced in that preparation process tailor the surface properties and alter the active sites for adsorbate binding and catalytic processes.29-31 The impact of defects would be regulated by co-existing surface water in a realistic environment. Early experiment results demonstrated that the adsorption of CO2 was blocked by predosed H2O in the case of a defective TiO2(110) surface,32 which was inconsistent with the strengthened adsorption stability for CO2 through hydrogen bonding with H2O predicted by density functional theory calculations.33 These observations revealed that the reciprocal action for vacancy and surface water is complex and it is necessary to explore their interdependence within this process systematically. We recently suggested that the oxygen vacancy significantly influenced the


4

Page 5 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

adsorption behavior and favored the dissociation process of CO2 on a Zn2GeO4 catalyst.34 Understanding whether the two mutually promote or inhibit the dissociation process and if this reciprocal action is sensitive to oxygen defect type will be extremely significant to the CO2 reduction process. It is reported that the Zn2GeO4 nanoribbons with exposing {010} facets efficiently promoted the CO2 methanation reaction.25 Besides, it is observed that the thermodynamically more stable (010) surface exhibits higher activity than (001) surface toward activation of H2O, which being a determining step in the couple reaction between CO2 and H2O.35, 36 Therefore, We extended our previous study concerning on a systematic investigation of the effect of H2O adsorption features and the reciprocal action with oxygen vacancy in CO2 dissociation on a Zn2GeO4 (010) surface. We demonstrated that the CO2 dissociation process was substantially facilitated by the introduction of H2O on a perfect surface, which proceeded by a direct, H-assisted dissociation pathway according to the appearance of surface water. The reciprocal action for adsorbed H2O and vacancy was strongly sensitive to vacancy type and favored the direct dissociation process. Specifically, the introduction of H2O substantially decreased the reaction energy but increased the barrier for CO2 dissociation on hydrated O2c1- and O3c2-defect surface. However, H2O addition had a minimal effect on the reaction potentials on a hydrated O2c2-defect surface. We characterized the effect of water and vacancy on the fundamental mechanism of C–O bond scission during CO2 reduction to facilitate the development of high-efficiency catalysts. COMPUTATIONAL METHODS All periodic density functional theory (DFT) calculations were performed with the Cambridge Sequential Total Energy Package (CASTEP) code,37, 38 using ultrasoft potentials to accommodate the interaction between valence electrons and ion cores.39 Based on the test in section 1 in


5

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 51

supporting information, we used the Perdew’s and Wang’s 1991 (PW91) parameterized generalized gradient approximation (GGA) to describe the exchange correction energy.40, 41 The electronic wave functions were expanded on a plane basis set with a cutoff energy of 340 eV. The Monkhorst-Pack mesh 3×3×3 for Zn2GeO4 bulk and 2×1×1 for the (010) surface was chosen for integration over the first Brillouin-zone. The maximum forces on an atom in any direction were converged within 0.05 eV/ Å to guarantee the structure optimization. The tolerance for energy, maximum displacement, and maximum stress was set at 2.0×10−5 eV/atom, 0.002 Å, and 0.1 GPa, respectively. The ternary oxide Zn2GeO4 acted as a semiconductor and crystallized in a spinel-like structure. The bulk lattice parameters of calculated equilibrium structures (a=b=14.63 Å, c=9.67 Å) were closely aligned with experimental values (a=b=14.23 Å, c=9.53 Å) with 3% deviation.42 The relaxed, dry, perfect Zn2GeO4 (010) surface (Figure 1; DP) was modeled by exposed threefold Zn-termination (Zn3c), threefold Ge-termination (Ge3c), threefold O-termination (O3c), and twofold O-termination (O2c). Seven layers were applied to distribute the 84 atoms and yielded a lattice of 9.67×14.63×20.29 Å. The interactions between periodic slabs were avoided by setting the vacuum region to 12 Å. The adsorbates and three topmost layers were allowed to relax in the calculations, while the other layers remained fixed. The humidity of the catalysts determines the features and stability of water on the surface, so dissociative and molecular adsorption motifs for water substantially simplified the water content on perfect Zn2GeO4 (010) surfaces.43, 44 For clarity, a HP was used to prefix the hydrated perfect (010) surface: HP-1 represents the H2O dissociatively adsorbed at Ge3c1…O2c2 site with an adsorption energy of -3.01 eV; HP-2a represents the bidentate-binding H2O molecule at Zn3c1…O2c1 site with an adsorption energy of -1.66 eV; and HP-2b represents the monodentate-


6

Page 7 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

binding H2O at the Zn3c4 site with an adsorption energy of -0.50 eV. H1 and H2 were used to mark the two hydrogen atoms of adsorbed H2O while Ow indicated the oxygen atom. In the HP-1 configuration, the O atom of H2O was coordinated with a Ge3c1 atom at a distance of 1.81 Å, while the H2 atom was coordinated with the O2c2 atom to form a surface hydroxyl. As a result, the dihydroxylated surface was produced when the distance of H2 and Ow was elongated to 1.77 Å. In the HP-2a configuration, H2O interacted with the surface through Ow–Zn3c1 and H2–O2c1 bonds. The favorable configuration for HP-2b is shown in Figure 1 in which the H2O was perpendicular to the surface and formed an Ow–Zn3c4 bond (2.25 Å). The defect (010) surface was created by removing a surface O2c or O3c atom and exposing the subsurface Zn and Ge atoms. Similarly, we used DD to denote the dry defect and HD to denote the hydrated defect surface. DDVo2c2 represented the dry (010) surface with an O2c2 vacancy, whereas HDVo3c2 refers the hydrated (010) surface with an O3c2 vacancy. We constructed hydrated defect surfaces by pre-adsorbing dissociative and molecular H2O on each dry defect surface to conceptualize the reciprocal roles of H2O and oxygen defects in the CO2 dissociation process. After comparing the CO2 dissociation process on a hydrated surface with the same oxygen vacancy (see Supporting Information), three surfaces with characteristic reactivity were obtained: H2O molecule bound at the Zn3c1…O3c1 site on the O2c1-defect surface in a bidentate state (HDVo2c1); the H2O molecule interacting with the O2c2-defect surface in a bidentate state at the Zn3c1…O2c1 site (HDVo2c2); H2O dissociative adsorption at the Ge3c1…O2c2 site on the O3c2defect surface. Energies and structural parameters are shown in Figure 1.


7

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 51

Figure 1 Top and side views of the dry and hydrated Zn2GeO4 (010) surface. Green, violet, and white balls represent Ge, Zn, and H atoms, respectively. The black dotted circles indicate the oxygen vacancies. The adsorbate O atoms are orange while surface O atoms are red. The adsorption structure stability was measured by the adsorption energy, which was defined as: Eads =Eslab+mol -Eslab -Emol In this formula, the Eslab+mol term was the total energy of a slab after adsorption where the Eslab term denoted as the energy of a bare slab, and Emol was the energy of gas phase molecule. The reaction barrier and energy were defined as: Ea =ETS -Eini ACS Paragon Plus Environment

8

Page 9 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Er =Efin -Eini where ETS was the energy of a transition state, Eini was the energy of an initial configuration, and Efin represented the energy of the final configuration. The complete LST/QST method was used to search the minimum energy paths of CO2 dissociation.45 To confirm the transition states, we used climbing image nudged elastic band method (CINEB) in Vienna ab initio simulation package (VASP) to search the minimum energy paths (in supporting information).

46

It was

observed that there was a certain gap in the obtained barrier values between two methods. However, the effect of catalyst structure and environmental on the dissociation process remained invariable. Moreover, from the viewpoint of geometries of the transition states, the bonding forms were basically the same. Considering our calculations based on CASTEP code, the results obtained by LST/QST were chosen as transition states in our work.

RESULTS AND DISCUSSION Effect of preadsorbed H2O on CO2 adsorption behavior on perfect and defect Zn2GeO4 (010) surfaces The adsorption of CO2 on the catalyst surface is a pivotal step of the reduction process and is significant to conversion engineering. We explored the dependence of CO2 adsorption behavior on the surface structure and environment. Our calculations showed that irrespective of surface state, CO2 interacted with the surface by bending its structure to the most stable configuration. This was also evident from previous experimental and theoretical results.47 Therefore, we determined the most favorable configurations for CO2 adsorption and used them as initial states for reactants. Oa and Ob were used to mark the two oxygen atoms of adsorbed CO2 which were


9

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 51

not in an equivalent position. We first considered the location of CO2 on a dry and H2O-covered perfect Zn2GeO4 (010) surface to obtain the most stable adsorption type as shown in Figure 2. The Oa atom of CO2 was attached to the surface metal site and the C atom pointed to the surface O atom, which created the COb slant on the perfect surface. The CO2 preferred to locate at the Ge3c1…O2c2 site on the dry perfect surface and formed Oa-Ge3c1 (1.86 Å) and C-O2c2 (1.44 Å) bonds that were exothermic by 1.73 eV in the DP-mol configuration. The effect of H2O introduction on adsorption geometries and energies for CO2 was closely related to the preadsorption features. Since the H2O molecule was dissociatively preadsorbed at the Ge3c1…O2c2 site, the secondarily stable Zn3c2…O2c3 was a priority for CO2 adsorption. The longer distance of Oa–Zn3c2 (1.96 Å) and C–O2c3 (1.48 Å) bonds formed in the HP-1-mol configuration was related to a higher adsorption energy (-0.81 eV) relative to the DP-mol configuration (-1.73 eV). Unlike the HP-1-mol configuration, the CO2 interacted with the H2O molecule-covered surface by forming Oa–Ge3c1 (1.92 Å and 1.83 Å) and C–O2c2 (1.47 Å and 1.42 Å) bonds in the HP-2a-mol and HP-2b-mol configurations. The location position and shortened bond lengths formed in HP-2a-mol and HP-2b-mol relative to HP-1-mol resulted in a stronger interaction between the CO2 and H2O molecule-covered surface, with adsorption energies of -1.20 eV and -1.84 eV, respectively. The raised steric hindrance for CO2 in the HP2a-mol was responsible for the reduced stability when compared to the DP-mol configuration because the preadsorbed bidentate H2O molecule was located obliquely over the Ge3c1 atom. There was an inverse trend in the adsorption of CO2 on the monodentate H2O-covered surface with regard to adsorption energy. As shown by the HP-2b-mol configuration (Figure 2), the monodentate H2O affected the configuration by slightly decreasing the corresponding bond


10

Page 11 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

lengths for CO2 interaction with the surface and was concomitant with decreased adsorption energy.

Figure 2 Top and side views of CO2 adsorption on the dry and hydrated perfect Zn2GeO4 (010) surface. Color coding is given in Figure 1, with the O atom of CO2 in orange and C atom in gray. As stated above, the CO2 molecule readily adsorbed on a surface that exposed preferable Ge3c1…O2c2 sites. These sites were shielded by fewer atoms and allowed close interaction with the surface through O-Ge3c1 and C-O2c2 bonds while avoiding an energy cost due to the steric


11

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 51

effect. In the case of a hydrated surface, the presence of dissociative H2O greatly decreased the CO2 adsorption stability because it occupied the most advantageous location available. When an H2O molecule was introduced on a perfect surface, bidentate H2O decreased the stability of CO2 adsorption by shielding the Ge3c1 atom, while shorter bonds lengths strengthened CO2 stability on a surface with preadsorbed monodentate H2O. Consequently, CO2 adsorption was sensitive to the surface environment with increasing stability in the following order: dissociated H2O < bidentate-binding H2O < dry surface < monodentate-binding H2O. Next, we will discuss CO2 adsorption on the defect surface. As shown in Figure 3, there were two locations for favorable adsorption: the vacancy and the vicinity. For the DDVo2c2 and DDVo3c2 surface, the CO2 interacted with the surface by filling its O atom in the vacancy and the C atom bonded to the surface metal. This configuration produced the COa(b) perpendicular to the surface with an adsorption energy of -1.46 eV and -1.56 eV. Such adsorption configurations (DDVo2c2mol, DDVo3c2-mol) were less stable than that on the perfect surface (DP-mol, -1.73 eV). The surface structure in Figure 1 accounted for this phenomenon. Both of the metal atoms surrounding the vacancy were buried under the surfaces, which resulted in greater difficulty of the CO2 molecule attachment due to higher steric hindrance. Unlike the DDVo2c2-mol and DDVo3c2-mol configuration where the oxygen vacancy was preferred, the nearby Ge3c1…O2c2 was the preferable site for CO2 adsorption in the DDVo2c1-mol configuration with adsorption energy of -1.97 eV. The elevation for Zn3c1 atom was low on the O2c1-defect surface which would direct the CO2 away from the oxygen vacancy. Regarding the adsorption energy, the introduction of H2O decreased the ability for CO2 to adsorb on the defect surface by increasing the binding distance. Moreover, the preadsorbed H2O had the greatest negative effect on CO2 adsorption in the O3c2 vacancy. As shown in Figure 3 HDVo3c2-mol configuration, the H2O-covered Ge3c1 site prevented


12

Page 13 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

CO2 attachment with the greatest decrease in adsorption ability by 0.82 eV.

Figure 3 Top and side views of CO2 adsorption on the dry and hydrated defect Zn2GeO4 (010) surface. Color coding is given in Figure 1, with the O atom of CO2 in orange and C atom in gray. The above results indicated that the CO2 adsorption was surface-state sensitive, with adsorption energies varying from -1.97 eV to -0.74 eV. The CO2 adsorption at the oxygen vacancy site was less stable than that on a perfect surface, and the generation of O2c1-defect made the vicinity site more preferable as a CO2 anchor. Moreover, the steric hindrance and variation of bonding state induced by preadsorbed H2O suppressed the adsorption for CO2 when the defect surface was hydrated. Direct CO2 Dissociation Direct dissociation and H-assisted dissociation are two general pathways of CO2 activation. As


13

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 51

accepted in the literature, the H released from H2O would participate in the scission process within a water-containing system.48, 49 Therefore, the combined effect of H2O and oxygen defects on the adjustment of the CO2 activation potentials and the selectivity of the activation mechanism are discussed in next section. The influence of the surface state on CO2 direct dissociation was described using the most favorable adsorption configurations for CO2 and by separating the CO and O fragments as the initial state and final state. As indicated by the adsorption energies in Table 1, the adsorption configurations for CO and O fragments on an H2O-covered surface were more stable than those on a dry surface, which would stabilize the dissociation process. There were significant differences in the adsorption geometries (Figure S1) that depended on the interactions of detached O and CO species with the hydrated surface. Compared to CO and O both located at the bridge surface site and shared Ge3c1 atom on a dry surface, the O atom adsorbed at the bridge site and bound to dissociative H2O on the HP-1 surface decreasing the adsorption energy by 0.40 eV. The hydrogen bond between the detached O atom adsorbed at the top site and bidentate binding H2O created the configuration of the dissociated product on HP-2a that was most stable with adsorption energy of -2.97 eV. The dissociation product on the HP-1 surface was less stable than that on the HP-2a surface due to the energy cost of preadsorbed dissociative H2O changes to the molecular configuration (Figure S1(b)). Although the situation for CO and O on the HP-2b surface was similar to that on DP surface, the presence of monodentate H2O shortened the formed bond length and lowered the adsorption energy by 0.15 eV. Besides the CO2 molecule, the adsorption behavior of the dissociative product also depended on the surface environment. We addressed the direct dissociation process on the perfect surface from the structures of reactant and dissociation product adsorption. Key parameters for energies and transition state


14

Page 15 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

geometries are given in Table 1 and Figure 4. The obtained reaction barrier and energy of C–O bond scission were high at 3.72 eV and 3.16 eV, respectively for the breakage step on the DP surface. In the configuration shown in Figure 4(a), the surface Ge3c1 atom was shared by CO and O fragments, which caused site competition and repulsion that substantially raised the energy cost. Table 1 shows that the C–O bond scission experienced lower barrier and reaction energies on a dissociative (HP-1) and bidentate H2O–covered (HP-2a) surface than on a dry (DP) surface, but the potentials were not altered by monodentate-binding H2O (HP-2b surface). This suggested that the process of CO2 dissociation was sensitive to the appearance of surface water. The introduction of dissociative H2O lowered the activation barrier and reaction energy of C–O bond breaking to 3.38 eV and 2.77 eV, respectively. The reason for this promotion was that no atom was shared for CO and O fragments located at the irrelevant site in the Figure 4(b) configuration, unlike the (a) configuration. The hydrogen bond formed between the Oa and preadsorbed H2O (1.77 Å) stabilized the transition state. When compared to the Figure 4(c) configuration, the deformation of dissociative H2O in the (b) configuration raised the energy cost. Therefore, the obtained barrier and reaction energies for breaking the C–O bond on the HP-1 surface were 0.47 eV higher than those on the HP-2a surface (Ea = 2.91 eV, Er = 2.30 eV). Regarding adsorption structures, the transition state of the C–O bond scission on a monodentate H2O-covered surface was similar to that on a dry surface, and both involved Ge3c1 site sharing. A similar energetic preference with a barrier of 3.79 eV and reaction energy of 3.12 eV was obtained for breaking the C–O bond on the HP-2b surface. The environmental sensitivity for the reactant and dissociated product adsorption determined the sensitivity of the scission step. For a hydrated surface, the introduction of dissociative and bidentate H2O dispersed the location sites for product adsorption and formed hydrogen bonds in


15

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 51

the final state. The steric effect was avoided, and the transition state was stabilized leading the process to be more kinetically and thermodynamically labile. The preadsorbed monodentate H2O had little effect on the dissociation process. The distances of C and Oa were stretched to 1.822.41 Å which indicated that the structures of transition states were close to those of dissociated products. Table 1 Calculated adsorption energies for CO2, dissociative product located at the favorable site; Activation barriers, reaction energies for direct scission step; C–O bond length in initial and transition state. CO*+O* 2* Surface ECO (eV) Ea (eV) Er (eV) dini (Å) dTS (Å) ads (eV) Eads perfect DP -1.73 -2.16 3.72 3.16 1.33 1.82 HP-1 -0.81 -2.56 3.38 2.77 1.30 2.41 HP-2a -1.20 -2.97 2.91 2.30 1.34 2.06 HP-2b -1.84 -2.31 3.79 3.12 1.33 1.97 defect DDVo2c1 -1.97 -5.17 2.30 0.39 1.36 1.31 DDVo2c2 -1.46 -4.79 1.28 0.61 1.46 2.58 DDVo3c2 -1.56 -5.04 1.18 0.01 1.49 2.52 HDVo2c1 -1.69 -5.45 3.51 -0.18 1.37 2.07 HDVo2c2 -0.94 -4.51 1.30 0.67 1.47 2.58 HDVo3c2 -0.74 -5.69 1.72 -0.29 1.41 1.80

Figure 4 Geometries of transition states for direct CO2 dissociation on perfect surface: (a) DP


16

Page 17 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

surface; (b) HP-1 surface; (c) HP-2a surface; (d) HP-2b surface. Surface state regulation of hydration and vacancy was predicted to appear on the catalyst surface concurrently. We systematically explored the reciprocal action for H2O and oxygen defects in the CO2 dissociation process. Firstly, the effect of a vacancy on CO2 direct scission was investigated. The structures for CO and O adsorption on a defect surface are shown in Figure S2. Our calculations revealed that the dissociative adsorption configuration was equivalent to CO captured by a perfect surface due to the hollow surface accommodation of an O atom. Regarding adsorption energy, the locations of products on DDVo2c1 (HDVo2c1) and DDVo3c2 (HDVo3c2) surfaces were more stable than those on the DDVo2c2 (HDVo2c2) surface, which was rationalized by the available Ge3c1…O2c2 site. For a dehydrated defect surface, the dissociated barriers varied from 1.18–2.30 eV and reaction energies varied in magnitude by 0.61 eV, which implied that the direct dissociation step on the defect surface was energetically more favorable than that on the dry, perfect surface. The transition state structures can explain this phenomenon. From Figure 5(a)-(c), it can be seen that there no atom is shared by dissociated species because the O atom filled the vacancy and resulted in less spatial resistance than that of the perfect surface. Among these, the obtained barrier for CO2 dissociation on DDVo2c1 was high at 2.30 eV with modest reaction energy (0.39 eV). Conversely, for CO2 scission on the DDVo2c2 surface, the obtained barrier was modest (1.28 eV) and reaction energy was high (0.61 eV). For the DDVo3c2 surface, the barrier was the lowest (1.18 eV), and the reaction energy was only 0.01 eV. This structure-sensitivity of the CO2 dissociation reaction was the product of coordination between the CO2 molecular adsorption state and the dissociative adsorption configuration. As shown in the Figure 5(a) configuration, the dissociated COb twisted its geometry from slant form in the DDVo2c1-mol configuration to approach a vertical state, which raised the energy cost. However,


17

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 51

the COb was perpendicular to the surface in the DDVo2c2-mol and DDVo3c2-mol configurations, which facilitated the CO2 direct scission. A comparison of the migration for the COb from Oa to metal-oxygen site demonstrated that the distance on the DDVo3c2 surface was shorter than that on the DDVo2c2 surface. This created the kinetic favorability of the CO2 direct scission on the DDVo3c2 surface.

Figure 5 Geometries of transition states for direct CO2 dissociation on the defect surface: (a) DDVo2c1 surface; (b) DDVo2c2 surface; (c) DDVo3c2 surface; (d) HDVo2c1 surface; (e) HDVo2c2 surface; (f) HDVo3c2 surface. As with the case of a perfect surface, we calculated the CO2 dissociation process on a hydrated defective surface with oxygen vacancies (O2c1, O2c2, and O3c2) surrounded by dissociative and molecular H2O. A detailed discussion is presented in Section 4 of the supporting information. The CO2 direct dissociation process was more inclined to occur on an O2c1-defect surface with a


18

Page 19 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

preadsorbed bidentate H2O molecule, an O2c2-defect surface with a preadsorbed bidentate H2O, and an O3c2-defect surface with a preadsorbed dissociative H2O. Therefore, we discussed the reciprocal action for H2O and the oxygen defect in the CO2 dissociation process on the most active, hydrated, defective surface. Key energies and geometry parameters are shown in Table 1 and Figure 5. The dissociative adsorption configuration for CO2 on the hydrated defect surface was equivalent to a CO located on a hydrated perfect surface (see Figure S2). For the HDVo2c1 and HDVo3c2 surface, the hydrogen bond between H2O and dissociated Oa(COb) substantially contributed to the adsorption stability because the adsorption energies decreased to -5.45 and 5.69 eV, respectively. Consequently, the decreased stability for reactant adsorption and the increased stability of product adsorption on the HDVo2c1 and HDVo3c2 surfaces provided a greater driving force for CO2 direct dissociation. This exchange resulted in a lower reaction energy than on a dry surface (-0.18 eV and -0.29 eV respectively) and converted the endoergic reaction to exoergic. The calculated reaction barriers for CO2 scission increased to 3.51 eV and 1.72 eV respectively, which decreased the kinetic favorability of the process and reduced the reaction rate. The transition states are displayed in Figure 5(d)-(f) to intuitively illustrate the impact. In the Figure 5(d) configuration, although the hydrogen bond between bidentate H2O and Oa stabilized the structure, the subsurface Ge atom created a greater steric effect for dissociated Oa atom to bind than the Ge3c1 atom. This steric effect explained the higher activation energy on the HDVo2c1 surface. In the Figure 5(f) configuration, the surface Ge3c1 site competition for CO and dissociative H2O also prevailed over the stabilization of the hydrogen bond and raised the energy cost. Notably, there was no adsorption site move or site competition for products on the HDVo2c1 and HDVo3c2 surface (see Figure S2), the formed hydrogen bond in the dissociative state was the main driving force that thermodynamically promoted the process. A special situation was


19

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 51

observed on the HDVo2c2 surface. From Table 1, the introduction of H2O concurrently destabilized the adsorption configuration for CO2 and the CO and O fragments on the O2c2-defect surface. The similar transition state structures resulted in a comparable barrier and reaction energy for CO2 dissociation on the DDVo2c2 surface (Figure 5(e) configuration). Our results indicated that the CO2 direct dissociation process was highly dependent on the surface structure and was facilitated by the presence of an oxygen vacancy. The introduction of H2O on a defect surface may play a different role in the CO2 dissociation process by the foundation of the vacancy type. Although the presence of H2O on the O2c1- and O3c2- defect surface induced steric hindrance in the transition state and increased the activation energy, the binding strength of the dissociated product was enhanced by hydrogen bond formation enough to favor the dissociation process thermodynamically. The presence of H2O on the O2c2- defect surface had a negligible effect on CO2 direct scission both kinetically and thermodynamically. H-Assisted CO2 Dissociation via COOH It was assumed that the CO2 activation proceeded via H-assisted C–O bond scission by binding at the C-end or O-end to form HCOO or COOH species. However, in our calculations, the C atom of CO2 pointed to and was surrounded by the surface O atom in its adsorption configuration, suggesting the HCOO pathway might be unlikely.30 Therefore, only the C–O bond cleavage via a COOH intermediate was considered in our results.


20

Page 21 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Table 2 Calculated adsorption energies for involved species located at the favorable site via the COOH pathway. Surface HP-1 HP-2a HP-2b HDVo2c1 HDVo2c2 HDVo3c2

CO *+H2 O*

Eads 2

(eV) ECOOH*+OH* (eV) ads perfect -3.65 -8.88 -2.33 -6.19 -2.34 -5.81 defect -2.04 -6.84 -2.33 -7.03 -2.62 -7.18

ECOOH* (eV) ads

ECO*+OH* (eV) ads

-3.56 -3.43 -3.36

-3.21 -2.27 -3.62

-4.41 -2.52 -2.86

-4.30 -2.24 -1.84

We first explored the H-assisted C-O bond scission on a perfect surface. This process included the formation and dissociation of COOH*. The most favorable adsorption configurations for species involved in the elementary steps were used as the initial states and final states. For CO2*+H2O*→COOH*+OH*, we considered the various possible generation modes as shown in Section 5 of the supporting information. It was found that cis-COOaH(1)*, trans-COObH(2)*, and cis-COOaH(2)* species were the dominant products on HP-1, HP-2a, and HP-2b surface, respectively. The key energies and structure parameters for the most energetically favorable pathways on each hydrated perfect surface are presented in Table 2, Table 3, and Figure 6, respectively. Regarding adsorption energy in Table 2, it was found that the co-adsorption of CO2 and dissociative H2O with formed COOH and OH exhibited distinctive stability over the other surfaces. As a result, the reaction for CO2 coupled with dissociative H2O had a lower barrier (0.63 eV) and reaction energy (0.15 eV) than the reaction between the CO2 and H2O molecule due to the transition state geometry. As shown in Figure 6(b), the H1 move to the Oa atom occurred concurrently with the Ow-H2 bond formation to facilitate the hydrogenation. There were two strong hydrogen bond interactions between H1 and the adjacent OwH2 and CO2. Consequently, the bridge built by the H1 atom allowed the adsorbate atoms (H1, H2, Ow, Oa, and


21

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 51

C) and surface atoms (O2c2, O2c3, and Znsub2) to form an eight-membered ring that enhanced the stability of this transition state. Conversely, the distance that H1 (or H2) atom transferred from the Ow atom to Oa (Ob) atom in Figure 6(e) and (h) significantly increased the barrier for CO2 hydrogenation by remaining molecular H2O. For the coupled reaction on another hydrated surface, the energetic difference for reactant adsorption was not particularly large (0.01 eV; Table 2) and the higher stability of the hydrogenation product was the driving force for coupling CO2 with bidentate binding H2O. In the Figure 6(e) configuration, as the H2 atom left the H2O molecule and approached the Ob atom, it acted as a medium to connect the isolated OH fragment with CO2 and stabilized the transition state. As a result, the barrier and reaction energy for CO2 coupled with bidentate H2O were 1.62 eV and 0.30 eV lower and more energetically stable in comparison to the coupled reaction for CO2 with monodentate H2O. In contrast, the barrier and reaction energy for CO2 coupled with monodentate H2O were as high as 4.62 eV and 1.91 eV, respectively. Such unfavorable potentials for hydrogenation could be rationalized by distance that H2 atom deviated from Ow atom to the CO2 in transition state (Figure 6(h)). Also, the adsorption of COOH and OH species was probably not stale enough, and so, the driving force was not powerful. Therefore, the reactivity for CO2 coupled with preadsorbed H2O was sensitive to the appearance, with dissociative H2O being most active species to donate an H atom.


22

Page 23 of 51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Table 3 Calculated activation barriers, reaction energies for the elementary step in H-assisted CO2 dissociation. Length for undergoing formation (hydrogenation) and cleavage (dissociation) bond between atoms in initial state and transition state. CO2*+H2O* → COOH*+OH* d1 Ea 1 Er 1 (Å) (eV) (eV)

Ea 2 (eV)

HP-1 HP-2a HP-2b

0.63 3.00 4.62

0.15 1.61 1.91

1.19 1.32 2.40

2.05 3.63 1.80

HDVo2c1 HDVo2c2 HDVo3c2

4.46 1.99 0.80

0.45 0.56 0.52

2.59 1.23 1.29

3.84 1.43 2.63

Surface

COOH* → CO*+OH* dCOOH Er 2 (Å) (eV) perfect 1.56 1.42 2.60 1.32 0.50 1.43 defect 1.44 1.29 0.74 1.38 2.24 1.33

d2 (Å)

CO2*+H2O* → CO*+2OH* Er Ea (eV) (eV)

2.79 2.42 2.41

2.20 5.24 4.62

1.83 4.21 2.41

2.47 1.86 2.33

4.46 1.99 3.15

1.89 2.30 2.76

Figure 6 Geometries for (a)(d)(g) COOH intermediates adsorption, (b)(e)(h) the transition states


23

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 51

for COOH formation, and (c)(f)(i) transition states for COOH dissociation. Structures for (a)-(c) were obtained on the HP-1 facet, (d)-(f) were obtained on the HP-2a facet, and (g)-(i) were obtained on the HP-2b facet. For COOH*→CO*+OH*, the adsorption for COOH in different configurations differed by

Critical Role of Water and Oxygen Defects in CâO Scission during

Recommend Documents

Critical Role of Water and Oxygen Defects in CâO Scission during